Difference between OSEN and OWE for secure open Wi-Fi

Question

I would like to better understand the difference and trade-offs between WiFi's OSEN and OWE authentication modes. My understanding is that these have broadly similar goals, namely preventing traffic from going out unencrypted over the air while simultaneously not requiring user credentials (i.e. being essentially an open network). Now, if I understand correctly

1. OSEN is essentially just WPA2 with EAP-TLS except that the client doesn't send any credentials. This still helps encryption, because the WPA2 MSK is derived from the TLS key, so the TLS key exchange is used to bootstrap the symmetric key for the subsequent communication.
2. OWE is essentially just WPA3 with a DH key exchange up front

Is this understanding correct? If so, what are the implications for deployment. I found very little actual documentation on OSEN, so I assume it may not be particularly widely implemented in clients? Are there any other practical considerations that set the two apart?

Answer 1

The main difference is that in OSEN, server's identity is authenticated using Anonymous EAP-TLS (and DAA), and still protecting the client's identity (identity is generated based on a private key which is part of the TPM HW module). See https://www.commscope.com/globalassets/digizuite/1527-wp-hotspot-2-0.pdf for an overview and https://en.wikipedia.org/wiki/Direct_Anonymous_Attestation for details on how TPM is used (DAA). As this is mainly for server identity, it's usage is limited for provisioning only, and disallowed during operation, typically using a separate SSID.

In OWE, no authentication is provided, an attacker can impersonate an AP successfully. It's main purpose is to provide security after the exchange.